1992 Miata MS3 Timing Retard with RPM

[acedeuce802]

I'm helping my friend out with his MS3+MS3x installation. This is on a 1992 Mazda Miata with all factory trigger sensors. The vehicle has an M45 supercharger, and LS1 coils, but is stock otherwise (other than removing emissions and other things MS doesn't use). The car feels like it has a spark retard rev limiter at 2000 rpm. We locked timing at 10 degrees, got out the timing light, and confirmed that it retards significantly with RPM. If I had to guess, I'd say 10+ degrees every 500 rpm. Changing the spark latency had little to no effect on this, but it seems that is for dialing in small amounts of timing scatter, and there is likely another underlying issue. We checked the resistor pots and retuned them. 2.48V on R11, 2.35V on R56 and R52 completely counter clockwise. Switching the timing prediction to no prediction allowed the engine to rev to around 4000 rpm. We took a composite log to prove that there is no sync loss. There is an msq and msl attached. Also there is a LED that is constantly lit and one that illuminates when I tip into the throttle, but we're not sure what those are for.

[elutionsdesign]

Try Ignition input capture: Falling edge. re-sync timing and see if it's steady.

[acedeuce802]

Using falling edge won't allow timing sync, and the engine doesn't start.

[jsmcortina]

What exactly happens? What does a strobe show?

James

[acedeuce802]

It feels and sounds just as if we had set a rev limiter at 4000 rpm, starting spark retard at 3500. When using a timing gun, and locked at 10 degrees advance, the timing mark shoots to the right upon tip in, then is too far out of sight to see how retarded it actually is. This was done before we changed the ignition settings and got it to rev to 4k, when it was only revving to 2k. As the RPM's drop, the timing mark becomes visible at maybe 1300 rpm (just guessing based off sound), where it jumps right to about 10 degrees after TDC, then as the motor lowers from 1300 to 800 RPM, the timing mark linearly moves from 10 degrees ATDC to 10 degrees BTDC.

Now that it revs to 4k, we can try using the timing gun again to see how off the timing is currently.

[jsmcortina]

Sorry, I was meaning what happens when you flip the ignition input capture? That was going to be my suggestion too.

Do remember that dial-back timing lights are inaccurate under transient conditions. A dumb light usually tells the truth.

James

[acedeuce802]

Ah, sorry. When going to falling edge (from what I remember), the tach (in Tunerstudio) jumps up to exactly 200 rpm momentarily, then immediately back to zero, while cranking. The time that it spends at zero is rougly twice the amount of time it spends at 200 rpm. Looking at the sync graphs shows very inconsistent cam and crank inputs, and lots of sync loss. I can get a composite and full log of that tonight.

Thanks for the help.

[acedeuce802]

Attached are current logs now that the engine will rev to 4000 rpm. We tried taking a composite log after changing the ignition signal to falling edge, but it produced no signal, and didn't create a log. The rpm jumped to 100 rpm every second or two, but was zero for the majority of the cranking period.

Just realized I don't have the .msq on this computer, the only change from the original one posted is that it uses the timing table, instead of fixed advance, and there's no timing prediction used.

[acedeuce802]

Bump, any ideas? Still having the same issue, and haven't been able to come up with any conclusions yet.

[Matt Cramer]

It is not losing sync. Does it break up under free revving too?

Can you check if the timing behaves itself at 4000 RPM if you set it to fixed timing?

[elutionsdesign]

With the fixed timing set to whatever you want, 10 preferably, it should stay at 10 regardless of rpm. Does it do that? Your timing table is not good and if the wideband is reading correctly its drowning in fuel as well. All of that is secondary to the timing being constant and CORRECT.

[billr]

I see coil dwell is 5msec, and LS1 coils being used; are those the kind of "LS coils" that can't take over-dwell?

[acedeuce802]

The dwell was updated when the LS coils were installed, the map was a few days old.

The issue is fixed, we just had to add the 1K pull-up resistor. It surprises me that there was no sync loss. And to answer the previous question, even with fixed timing, it was still retarding like crazy. But it now holds 10 degrees fixed timing to redline.

Thanks for the help.